SEARCH

SEARCH BY CITATION

Keywords:

  • Growth;
  • insulin-like growth factor;
  • nutrition;
  • small-for-gestational age

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

Please cite this paper as: Jones R, Cederberg H, Wheeler S, Poston L, Hutchinson C, Seed P, Oliver R, Baker P. Relationship between maternal growth, infant birthweight and nutrient partitioning in teenage pregnancies. BJOG 2010;117:200–211.

Objective  Teenagers are susceptible to delivering small-for-gestational-age (SGA) infants. Previous studies suggest that maternal growth may contribute, as a result of preferential nutrient partitioning to the mother. We investigated the impact of maternal growth on birthweight in pregnant teenagers in the UK, and examined endocrine mediators of nutrient partitioning.

Design  A prospective observational multicentre study, About Teenage Eating, conducted between 2004 and 2007.

Setting  Four hospitals in socially-deprived areas of Manchester and London.

Population  A total of 500 pregnant adolescents (14–18 years of age) with a singleton pregnancy were recruited at 10–21 weeks of gestation, with follow-up studies on 368 subjects. A cohort of 80 pregnant adults (25–40 years of age) provided a control group for determining growth.

Methods  Skeletal growth, weight gain and skinfold thickness were measured from first to third trimester, together with maternal levels of micronutrients and metabolic hormones: insulin-like growth factor (IGF) system and leptin. Dietary analyses were performed.

Main outcome measure  SGA birth.

Results  Maternal growth was not associated with SGA birth: growing mothers delivered more large-for-gestational-age infants (OR 2.51; < 0.05). Growers had greater weight gain (< 0.001), fat accrual (< 0.001) and red cell folate concentrations (< 0.01) than non-growers. Maternal IGF-I (< 0.01) and leptin (< 0.001) were positively associated with maternal and fetal growth, whereas IGF-I (< 0.001) was negatively associated. Teenagers that were underweight at booking or with low weight gain were at greater risk of SGA birth.

Conclusions  Maternal growth was not detrimental to fetal growth in this UK population of teenagers. Greater weight gain and higher concentrations of IGF-I in growing teenagers may provide anabolic drive for maternal and fetal growth.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

Teenage pregnancy is a prominent social problem in both the USA and the UK, where births to young women aged 13–19 years account for 10% and 7% of all births, respectively.1,2 Teenagers have higher rates of adverse obstetric outcomes, particularly the delivery of low birthweight or preterm infants.3–5 Socio-economic and lifestyle factors have been implicated, although adjustment for these confounding variables demonstrates a sustained elevated risk of poor pregnancy outcome.6,7

Gynaecological immaturity may contribute to this increased risk;6,8 however, continued maternal growth has also been implicated. Skeletal growth velocity peaks prior to menarche, but can continue at a slower rate into late adolescence, with wide variation between individuals.9–11 Although spinal lordosis and kurtosis mask gain in stature during pregnancy, alternative sensitive measures of growth have revealed the potential for continued growth in up to 50% of 13- to 18-year-old subjects.12,13

The impact of maternal growth on pregnancy outcome is, however, controversial. Studies of pregnant teenagers (predominantly Black and Hispanic) in Camden, NJ, USA, demonstrated that those with > 1 mm of growth in knee height over 6 months delivered infants that were 150–200 g lighter than non-growers, and experienced a higher rate of preterm delivery.12,14,15 These effects were more pronounced in the younger or multiparous gravida. It was hypothesised that this was the result of preferential partitioning of nutrients to the growing mother, as growing teenagers gained more weight in late pregnancy, to the detriment of the fetus.15 Similar findings from a retrospective analysis of National Collaborative Perinatal Project records from 1959 to 196516 and studies of adolescent sheep17 lend support to this theory. In contrast, Stevens-Simon et al.13 showed that Black American teenagers with immature bone age were not susceptibile to lower birthweight deliveries, nor was there evidence of abnormal nutrient partitioning.

These conflicting data highlight the need to investigate whether teenagers that continue to grow during pregnancy are a high-risk group for pregnancy complications, and whether metabolic adaptations that regulate nutrient partitioning are disrupted by maternal growth. In normal adult pregnancy, metabolic adaptations occur in the third trimester to prioritise transfer of dietary glucose and amino acids to the rapidly growing fetus. This switch from the anabolic state of early pregnancy is mediated by a pregnancy-specific hormonal milieu that promotes lipolysis of maternal adipose stores accrued in early pregnancy, and attenuates glucose uptake by maternal tissues, thereby increasing nutrient availability for the fetus.18 These adaptations are induced by placental hormones19,20 through their metabolic actions in inducing insulin resistance and stimulating the production of maternal insulin-like growth factor-I (IGF-I), a key mediator of lipolysis and nutrient transfer across the placenta.18,21–23 Whether metabolic adaptations to pregnancy are disrupted in growing pregnant teenagers has not been assessed, but imbalances in the growth hormone-IGF-I axis have been described in rapidly-growing pregnant adolescent sheep.24,25

The objective of the current study was to interrogate the relationship between maternal growth and small-for-gestational-age (SGA) birth in a cohort of teenagers living in the UK by anthropometric and nutritional analyses. Assessment of the metabolic biomarkers, leptin and IGF-I, together with binding proteins IGFBP-1 and -3, were undertaken to elucidate whether alterations in fat accrual and metabolic adaptation underpin the susceptibility of growing teenagers to adverse pregnancy outcomes.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

Study design and participants

The About Teenage Eating (ATE) study was a prospective observational study of pregnant teenagers living in socially-deprived areas of the UK. From November 2004 to February 2007, study-specific midwives recruited 500 pregnant teenagers from antenatal clinics at four hospitals in London and Manchester. Criteria for inclusion were: singleton pregnancy, 14–18 years of age and <21+0 weeks of gestation. Exclusion criteria included pre-existing medical or obstetric disorders, multiple gestation or history of three or more miscarriages. A cohort of pregnant adults (n = 80, 25–40 years of age), was recruited (n = 40 from London; n = 40 from Manchester) to act as a control group for growth assessment. Other than age, entry criteria were identical with those of the teenagers.

Demographic information, self-reported smoking status and recalled pre-pregnancy weight were obtained from all participants at recruitment. Gestational age was calculated from the last menstrual period and confirmed by ultrasound measurement. The study was approved by the Central Manchester Research Ethics Committee (#03/CM/032). Written informed consent was obtained from all participants.

Anthropometric assessment

Measurements were taken at recruitment (median 13+6, range 10–20 weeks of gestation) and at 28–32 (median 29+5) weeks of gestation by research midwives. Height was assessed using a metal stadiometer to the nearest 0.5 cm. Weight was measured to the nearest kg using standard hospital scales with subjects lightly dressed. Knee height (distance from the heel to suprapatella) was measured to the nearest 0.25 mm using a knemometer (Force Institute, Copenhagen, Denmark), with the knee at 60° flexion. Nine measurements were taken at each assessment. The previously-reported measurement error is ±0.25 mm in pregnant subjects. The assessment of change in knee height in a cohort of pregnant adults evaluated any increase in knee height related to pregnancy per se (e.g. resulting from oedema or fat accrual), so as to enable the accurate identification of teenagers undergoing skeletal growth during pregnancy. However, there is minimal fat accrual in the suprapatellar region between these stages of pregnancy.15,26,27 Peripheral and centrally-distributed subcutaneous fat were assessed by measurement of skinfold thickness at triceps and subscapular sites using Holtain calipers (Holtain Limited, Crosswell, UK). Arm circumference was measured with a non-stretch tape measure to the nearest mm. All measurements were taken in triplicate. Arm circumference and triceps skinfolds were used to calculate an upper arm fat estimate and upper arm fat-free mass estimate.28 All anthropometric changes were standardised to a 90-day period between measurements. This was calculated by dividing the change between visits 1 and 2 by the number of days between the visits, and multiplying it by 90.

Endocrine assessments

A 30-ml sample of venous blood was collected between 28 and 32 weeks of gestation. Subjects were asked to eat only lightly beforehand, although for ethical reasons were not asked to fast. Plasma and serum were prepared and aliquots were stored at −80°C until assay. Plasma IGF-I and IGFBP-3 were assayed using commercially available enzyme-labelled chemiluminescent immunometric Immulite® assays (DPC, Los Angeles, CA, USA). The mean intra-assay variability was 5.8% for IGF-I and 1.7% for IGFBP-3. Plasma IGFBP-1 was determined using a radioimmunoassay.29 Mean intra- and inter-assay variability was 4.6% and 7.0%, respectively. Serum leptin was assessed by ELISA (DRG Instruments, Marburg, Germany). The intra- and inter-assay variability was 6.9% and 8.7%, respectively. All assays were performed on the same day by a single operator.

Assessment of micronutrient status

Maternal concentrations of selected important micronutrients were measured in the same blood samples to determine whether nutritional status differed between growers and non-growers. Red cell folate and serum folate, total homocysteine, vitamin B12 and ferritin were measured by competitive enzyme immunoassay. Serum iron was measured by colourimetric analysis. Assay methods and precision are reported in detail elsewhere.30

Assessment of dietary intake

Maternal dietary intake during the third trimester was assessed by three multiple-pass 24-h dietary recall interviews, occurring on non-consecutive days. All interviews were conducted by trained interviewers, either in person or by telephone, as previously described.30 Photographs of foods commonly eaten in the UK were used to improve the estimation of portion size. These were provided to participants at the first 24-h dietary recall, and were retained for subsequent interviews by telephone. Mean daily nutrient intake was calculated by using the UK Nutrient Databank (Her Majesty’s Stationery Office, Norwich, UK). Dietary data were excluded if energy intake was >3 SDs from the mean (n = 1).

Obstetric outcome

Infant birthweight, sex and gestational age at delivery were obtained from patient records. Customised birthweight centiles were calculated using Gestation-Related Optimal Weight (grow) software (Gardosi J, Francis A. Customised Weight Centile Calculator – GROW-Centile v.5.12/6.2 2009. Gestation Network, http://www.gestation.net), which adjusts birthweight for gestational age, infant sex and maternal constitutional factors known to influence birthweight: maternal height and weight at time of booking to antenatal care, parity and ethnicity.31 This enabled the classification of infants as SGA (<10th centile of expected birthweight for that particular pregnancy), appropriate for gestational age (AGA; between 10th and 90th centiles) or large for gestational age (LGA, >90th centile). Customised birthweight z-scores based on standard deviations from the mean were calculated. Preterm birth (at <37 weeks of gestation) and pre-eclampsia (gestational hypertension with proteinuria)32 were recorded.

Statistical analysis

Subjects were anonymised at recruitment, and data were entered on a secure internet-based database (MedSciNet, Stockholm, Sweden). Statistical analyses were performed using Stata 9.2 (StataCorp LP, College Station, TX, USA). Simple and multiple regression analyses with robust standard errors were used to determine associations between pregnancy outcome, anthropometric variables and endocrine/nutritional biomarkers. Where data were not normally distributed, biomarker data were log-transformed and presented as geometric means with standard deviations. The geometric SD represents the typical ratio between an individual value and the overall geometric mean. Comparisons between outcome groups were presented as ratios of geometric means (ratiogm) with 95% confidence intervals. Normally-distributed data were not transformed, although generalised linear modelling with a log link allowed presentation as ratios of arithmetic means.33 Adjustments were made for potential confounding variables associated with SGA birth: ethnicity, smoking, chronological and gynaecological age, body mass index (BMI) and socio-economic status (measured by receipt of housing benefit and the Index of Multiple Deprivation, IMD, score for the area). Comparisons of nutrient intakes were adjusted for total energy intake by multiple linear regression.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

Of the 500 teenagers recruited, 129 did not attend a second visit for assessment of growth, and three were lost to follow-up. Analysis was therefore limited to 368 subjects: the demographic profile of these subjects is shown in Table 1. Maternal plasma samples were collected from 262 of these subjects for endocrine and nutritional assessments. Data for upper arm fat area and fat-free mass estimates were limited to 231 subjects because of the late introduction of arm circumference measurements. There were no differences in demographic profiles or obstetric outcome between populations used for growth, anthropometric or endocrine analyses, nor from the total population recruited.

Table 1.   Demographic details and obstetric outcomes of study participants, n = 368
Age at recruitment: median (range) 17.8 (14.0–18.9) years
14–15 years: n (%) 27 (7.3)
16 years: n (%) 48 (13.0)
17 years: n (%)137 (37.2)
18 years: n (%)156 (42.4)
Gynaecological age: median (range)  5.0 (1.0–9.6) years
Ethnicity
White: n (%)198 (53.8)
Black Afro-Caribbean: n (%)100 (27.2)
Mixed White-Black: n (%) 38 (10.3)
South Asian: n (%)  7 (1.9)
Other ethnic group: n (%) 25 (6.8)
Smoking, self-reported at recruitment
Smoking at recruitment: n (%)107 (29.1)
Ex-smoker postconception: n (%) 70 (19.0)
Ex-smoker preconception: n (%) 38 (10.3)
Non-smoker: n (%)153 (41.6)
Parity: n (%) nulliparous348 (94.6)
Recalled pre-pregnancy weight: median (range) 57 (32–117) kg
Pre-pregnancy BMI: median (range) 21.6 (13.3–41.6) kg/m2
Height at recruitment: median (range)163 (148–186) cm
Weight at recruitment: median (range) 60 (40–120) kg
BMI at recruitment: median (range) 22.7 (15.4–44.6) kg/m2
Underweight (≤19.0): n (%) 36 (9.8)
Normal (19.2–24.9): n (%)125 (58.4)
Overweight (25.0–29.9): n (%) 77 (20.9)
Obese (≥30.0): n (%) 40 (10.9)
Obstetric outcome
SGA (<10th centile): n (%) 63 (17.1)
LGA (>90th centile): n (%) 23 (6.3)
Preterm delivery: n (%) 26 (7.1)
Pre-eclampsia: n (%)  9 (2.4)

A high rate of SGA (17.1%) was observed, whereas frequencies of LGA (6.3%), preterm birth (7.0%) and pre-eclampsia (2.5%) were within the expected ranges for the UK population.

Maternal growth in adolescent pregnancy

There were significant differences in the patterns of knee-height change in 90 days between adolescents and adults (Figure 1A), with a greater frequency and magnitude of increase being apparent in teenagers: 1.35 ± 1.89 mm/90 days in teenagers versus 0.42 ± 1.22 mm/90 days in adults. The small degree of knee-height change in adults represents pregnancy-associated soft tissue changes, as well as inherent measurement error, and is consistent with previous reports.26,27 A threshold for probable growth was defined by the 90% confidence interval (≥ 2 mm/90 days) for change in knee-height growth in adults. A third of the cohort was defined as ‘growers’ on this basis. Analyses were performed using both this threshold and change in knee height in 90 days as a continuous variable. There was no difference in chronological or gynaecological age or ethnicity between growers and non-growers, nor any relationship between these variables and knee-height growth as a continuous variable.

image

Figure 1.  Incidence of maternal growth in pregnant teenagers and impact on pregnancy outcome. (A) Histograms illustrating the change in knee height after 90 days in adult and teenage populations. The dotted line denotes zero; the dashed line indicates the threshold used to define growth at 2 mm/90 days. (B) Incidence of adverse pregnancy outcome in growing and non-growing teenagers: logistic regressions with adjustment for smoking status, deprivation indices, ethnicity, and chronological and gynaecological age. †Current smokers excluded for LGA analysis; *≤ 0.05.

Download figure to PowerPoint

Smoking was strongly and inversely associated with maternal growth in teenagers: mean knee-height growth was 0.95 ± 0.18 mm/90 days in smokers compared with 1.38 ± 0.15 mm/90 days in non-smokers (< 0.001). Only 20% of smokers were growers, compared with 38% of non-smokers (OR of growth with smoking, 0.41; 95% CI 0.24, 0.70; < 0.001).

Impact of maternal growth on pregnancy outcome

Maternal growth was not associated with reduced fetal growth. Instead it was positively related to birthweight z-score, although this association was of marginal significance following adjustment for confounding variables (r = 0.238; 95% CI, 0.01, 0.49; = 0.058). The mean birthweight z-score was significantly higher in growers than in non-growers (−0.14 versus −0.40 SD; < 0.05), and the mean birthweight z-score for growers’ infants did not differ significantly from the normal adult reference population from which the grow centiles were derived.31 In contrast, infants of non-growing teenagers had a significantly lower mean birthweight z-score compared with that expected in a normal adult population (< 0.0001).

Using knee-height change in 90 days as a continuous variable, growth was weakly protective against SGA (OR 0.84 of SGA per mm of knee-height growth; 95% CI 0.70–1.01; = 0.058). Although the incidence of SGA birth did not differ significantly between growers and non-growers (13% versus 20%; = 0.182) (Figure 1B), growing teenagers delivered more LGA infants (>90th centile) (15% versus 7%; < 0.05). There was no difference in the incidence of preterm birth or pre-eclampsia between growers and non-growers.

Impact of maternal growth on weight gain and adiposity

Gestational weight gain and fat accrual between the end of the first and early third trimesters were compared between growing and non-growing teenagers (Table 2). Pre-pregnancy BMI values did not differ between groups, although at recruitment growers weighed more (< 0.05) and had higher BMIs (< 0.01). These differences increased at 28–32 weeks (< 0.001), with growers gaining 0.6 kg/week compared with 0.4 kg/week in non-growers (< 0.001). All indices of maternal adiposity were increased in growers compared with non-growers (< 0.01 at visit 1; < 0.001 at visit 2). Upper arm fat-free mass estimates did not differ between groups.

Table 2.   Differences in weight and anthropometric measures in growing and non-growing teenagers during pregnancy
 Growing n = 121 Mean (SD)Non-growing n = 241 Mean (SD)Difference (95% CI)P value
  1. Growth: ≥2-mm increase in knee height/90 days. Linear regressions with robust standard errors, with adjustment for smoking status, deprivation indices, ethnicity, and chronological and gynecological age. *≤ 0.05, **≤ 0.01, ***≤ 0.001. ††Lower number of subjects because of the late introduction of the arm circumference measurement.

Weight (kg)
Visit 166.18 (15.01)62.75 (13.73)3.43 (0.21, 6.65)<0.05*
Visit 275.57 (15.47)69.27 (13.60)6.30 (3.02, 9.58)<0.001***
Change Visit 1–2 (kg/90 days)7.70 (3.71)5.25 (3.49)2.44 (1.65, 3.24)<0.001***
BMI (kg/m2)
Visit 124.89 (5.03)23.48 (4.58)1.40 (0.34, 2.46)0.01**
Triceps skinfold thickness (mm)
Visit 117.08 (7.12)14.88 (7.09)2.20 (0.61, 3.80)<0.01**
Visit 219.87 (7.77)15.92 (6.88)3.95 (2.27, 5.63)<0.001***
Change visit 1–2 (mm/90 days)2.44 (3.53)0.87 (3.69)1.51 (0.59, 2.36)0.001***
Subscapular skinfold thickness (mm)
Visit 119.21 (8.55)16.27 (9.71)2.95 (0.93, 4.96)<0.01**
Visit 223.39 (10.13)18.61 (9.45)3.95 (2.56, 7.02)<0.001***
Change visit 1–2 (mm/90 days)3.49 (4.03)1.91 (4.03)1.51 (0.67, 2.48)0.001***
Upper arm fat area estimate (cm2)††(n = 73)††(n = 158)  
Visit 126.49 (14.50)20.77 (13.01)5.72 (1.84, 9.60)<0.01**
Visit 231.96 (16.41)22.28 (12.95)9.69 (5.43, 13.94)<0.001***
Change visit 1–2 (cm/90 days)4.77 (6.51)1.30 (5.57)3.47 (1.75, 5.20)<0.001***
Upper arm fat-free mass estimate (cm2)††(n = 73)††(n = 158)  
Visit 135.82 (10.10)35.47 (8.15)0.35 (−2.30, 2.99)0.79
Visit 234.41 (8.84)35.19 (7.89)−0.78 (−3.16, 1.60)0.52
Change visit 1–2 (cm/90 days)−1.21 (5.45)−0.12 (5.25)−1.09 (−2.59, 0.41)0.15

Knee-height growth positively correlated with maternal weight gain, with a 0.58 ± 0.09 kg gain in weight/90 days for every 1 mm increase in knee height/90 days (< 0.001). Similar positive associations were observed for other anthropometric measurements. Changes in each parameter for 1 mm of gain in knee height/90 days were: triceps skinfold 0.38 ± 0.10 mm/90 days; subscapular skinfold 0.43 ± 0.11 mm/90 days; upper arm fat 0.83 ± 0.24 cm2/90 days (≤ 0.001 in all cases). There was no association between knee-height growth and upper arm fat-free mass.

Impact of maternal weight and adiposity on adverse pregnancy outcome

As growing teenagers experience greater weight gain and fat accrual, the effect of these measures on infant birthweight was determined. Table 3 shows differences in gestational weight gain and adiposity between subjects with SGA and non-SGA deliveries. As expected, subjects underweight at recruitment were more likely to deliver an SGA infant (OR 3.18; 95% CI 1.38, 7.32; < 0.01) than those in the normal BMI range. Interestingly, subjects with a high BMI were also at increased risk of an SGA birth (OR 3.70; 95% CI 1.676, 8.171; < 0.001).

Table 3.   Change in anthropometric measurements (changes/90 days) over pregnancy as predictors of SGA
Change visit 1–2SGA (n = 64) Mean (SD)Non-SGA (= 304) Mean (SD)OR (95% CI)P value
  1. Logistic regressions with adjustment for smoking status, deprivation indices, ethnicity, and chronological and gynecological age. *P ≤ 0.05, **P ≤ 0.01. ††n = 231 (39 SGA and 291 non-SGA) because of the late introduction of arm circumference measurements.

Weight (kg/90 days)4.86 (4.85)6.49 (3.34)0.87 (0.79, 0.95)<0.01**
Triceps skinfold (mm/90 days)1.00 (4.18)1.46 (3.59)0.97 (0.89, 1.05)0.40
Subscapular skinfold (mm/90 days)1.45 (4.62)2.63 (3.98)0.93 (0.86, 1.00)0.05*
Sum of skinfold (mm/90 days)2.48 (7.22)4.04 (6.20)0.96 (0.91, 1.01)0.09
Upper arm fat area estimate (cm2/90 days)††1.21 (6.57)2.60 (5.95)0.96 (0.90, 1.02)0.19
Upper arm fat-free mass estimate (cm2/90 days)††−1.59 (6.12)−0.21 (5.15)0.95 (0.90, 1.02)0.14

Teenagers delivering SGA infants had lower gestational weight gain than those delivering non-SGA infants, irrespective of BMI at booking (4.86 kg/90 days, 0.38 kg/week, versus 6.49 kg/90 days, 0.50 kg/week), demonstrating a protective effect of weight gain against SGA (OR of SGA 0.87; 95% CI 0.79, 0.95, for every kg/90 days gained; < 0.01). Anthropometric analyses revealed that subjects delivering SGA infants tended to have lower fat accrual, although with the exception of change in subscapular skinfold thickness, these differences narrowly failed to meet statistical significance (Table 3). Teenagers delivering LGA infants gained more weight (8.14 versus 6.09 kg/90 days), demonstrating increased risk of LGA (OR 1.12; 95% CI 1.02, 1.22, per kg/90 days gained; < 0.01) (data not shown).

Endocrine profile of pregnant teenagers: impact of maternal growth and infant birthweight

No effect of time since eating was detected for any of the hormones studied, for example, for IGFBP-1 r = −0.01, = 0.87. Growing teenagers had significantly higher serum leptin (< 0.001) and plasma IGF-I (< 0.05), but lower IGFBP-1 concentrations (< 0.001) (Table 4), than non-growing teenagers. IGFBP-3 did not differ between groups. Bioavailable IGF-I, as estimated by comparing molar ratios of IGF-I and its binding proteins, was significantly higher in growing versus non-growing teenagers. A doubling in the plasma IGF-I concentrations was associated with a gain in knee height of 0.73 mm/90 days (95% CI 0.23, 0.122; < 0.01), in line with IGF-I actions in promoting skeletal growth, and a similar relationship was observed for bioavailable IGF-I (data not shown). Serum leptin and plasma IGF-I concentrations were positively associated with maternal BMI and skinfold thicknesses at all gestations (data not shown).

Table 4.   Differences in endocrine measurements at 28–32 weeks of gestation in growing and non-growing teenagers
 Growing (= 78) Geometric mean (SD)Non-growing (n = 176) Geometric mean (SD)Geometric ratio (95% CI)P value
  1. Molar ratios calculated for IGF-I and IGFBPs. Ratio of geometric mean values by multiple regression analysis adjusted for smoking status, socio-economic status, ethnicity, BMI, and chronological and gynaecological age. The geometric SD represents the typical ratio between an individual value and the overall geometric mean. *≤ 0.05, **≤ 0.01, ***≤ 0.001.

IGF-I (ng/ml)349.54 (1.38)301.29 (1.43)1.13 (1.02,1.26)<0.05*
IGFBP-1 (ng/ml)50.14 (1.92)72.92 (1.87)0.74 (0.63, 0.88)<0.001***
IGFBP-3 (μg/ml)5.45 (1.25)5.30 (1.26)1.03 (0.97, 1.11)0.34
Leptin (ng/ml)29.28 (2.00)18.88 (2.10)1.40 (1.19, 1.65)<0.001***
IGF-I : IGFBP-128.99 (2.32)16.56 (2.31)1.54 (1.22, 1.95)<0.001***
IGF-I : IGFBP-30.39 (1.53)0.34 (1.36)1.10 (1.00, 1.21)<0.05*

As expected, significant relationships were observed between maternal IGF-I and IGFBP-1 concentrations and the incidence of SGA (Table 5). IGF-I concentrations were 15% lower (< 0.05) in teenagers delivering SGA infants, whereas IGFBP-1 was 42% higher (< 0.01). This suggests a 41% reduction in bioavailable IGF-I (≤ 0.01). Although the IGFBP-3 concentration was not significantly different, the ratio of IGF-I : IGFBP-3 implied there was 18% less free IGF-I in teenagers delivering SGA infants (≤  0.01). The converse was observed for teenagers delivering LGA infants, with a 20% increase in total IGF-I and a 23% decrease in IGFBP-1 concentrations (< 0.05 for both), resulting in elevated bioavailable IGF-I (56% increase in IGF-I : IGFBP-1, ≤ 0.01, and a 15% increase in IGF-I : IGFBP-3, < 0.05) (data not shown).

Table 5.   Differences in endocrine measurements at 28–32 weeks of gestation in teenagers delivering SGA or non-SGA infants
 SGA (= 38) Geometric mean (SD)Non-SGA (= 224) Geometric mean (SD)Geometric ratio (95% CI)P value
  1. Molar ratios calculated for IGF-I and IGFBPs. Ratio of geometric mean values by multiple regression analysis adjusted for smoking status, socio-economic status, ethnicity, BMI, and chronological and gynaecological age. The geometric SD represents the typical ratio between an individual value and the overall geometric mean. *≤ 0.05, **≤ 0.01.

IGF-I (ng/ml)278.69 (1.42)325.36 (1.43)0.85 (0.73, 0.99)<0.05*
IGFBP-1 (ng/ml)83.19 (2.20)62.35 (1.86)1.42 (1.14, 1.77)<0.01**
IGFBP-3 (μg/ml)5.37 (1.27)5.34 (1.25)1.03 (0.93, 1.13)0.58
Leptin (ng/ml)20.09 (2.24)22.08 (2.12)0.86 (0.65, 1.13)0.28
IGF-I : IGFBP-113.08 (2.78)21.22 (2.29)0.62 (0.42, 0.82)<0.01**
IGF-I : IGFBP-30.32 (1.37)0.38 (1.37)0.82 (0.72, 0.94)<0.01**

A doubling in IGF-I was associated with an increase in the birthweight z-score of 0.32 SD, whereas a doubling in IGFBP-1 was associated with a reduction of 0.37 SD (< 0.001). Maternal leptin concentrations were not associated with birthweight z-score. IGF-I and IGFBP-3 concentrations were 15% and 16% higher, respectively, in teenagers who delivered preterm compared with those who delivered at term (< 0.05 for IGF-I; < 0.001 for IGFBP-3). The ratio of IGF-I : IGFBP-3 did not differ between preterm and term pregnancies. No differences were detected for any of the other endocrine parameters.

Nutritional status of growing and non-growing teenagers

Mean red cell folate concentrations were 15% higher in growers compared with non-growers (< 0.05; Table 6). A similar trend was observed for serum folate and vitamin B12, although these differences failed to reach statistical significance after adjustment for confounding variables. Total serum homocystine (tHcy) concentrations were higher in growing than non-growing teenagers, although this was of marginal significance (= 0.058). Serum iron concentrations did not differ between the cohorts.

Table 6.   Concentrations of nutritional biomarkers at 28–32 weeks of gestation in growing and non-growing teenagers
BiomarkerGrowing Geometric mean (SD)Non-growing Geometric mean (SD)Geometric ratio (95% CI)P value
  1. Ratio of geometric mean values by multiple regression analysis adjusted for smoking status, socio-economic status, ethnicity, BMI, and chronological and gynaecological age. The geometric SD represents the typical ratio between an individual value and the overall geometric mean. ≤ 0.05.

Serum folate (nmol/l)13.4 (3.7)     (n = 92)12.3 (3.9)    (n = 189)1.11 (0.98, 1.26)0.10
Red cell folate (nmol/l)701 (3.6)     (n = 86)618 (3.3)    (n = 171)1.15 (1.02, 1.28)0.017*
Serum tHcy (μmol/l) 7.5 (1.4)     (n = 92)8.1 (1.4)    (n = 191)0.93 (0.85, 1.00)0.06
Serum vitamin B12 (pmol/l)187 (1.1)     (n = 92)173 (1.2)    (n = 191)1.09 (0.98, 1.22)0.10

Subjects who provided complete dietary data (n = 278) did not differ in terms of demographic or biophysical characteristics from those who did not. There was no difference in the incidence of SGA between the groups; however, the rate of preterm birth was higher in those who did not provide dietary data (16.0% compared with 5.9%; < 0.001 by chi-square test). Nutrient intakes exceeded the UK Reference Nutrient Intake (RNI), with the exception of vitamin A, folate, magnesium, iron and vitamin D (reported in full in Baker et al.30). There were no differences in energy or macronutrient intake between growing and non-growing teenagers (Table 7). However, greater intake of several micronutrients, notably calcium, magnesium, phosphorus, iron and riboflavin, was detected in growing versus non-growing teenagers, providing evidence for better dietary quality.

Table 7.   Mean daily energy and nutrient intakes of growing and non-growing teenagers, excluding intake from supplements
 Grower (n = 91)Non-grower (n = 187)Comparison (95% CI)Fully-adjusted (95% CI)
  1. Macronutrients are presented as arithmetic means (SD) and micronutrients are presented as geometric means (SD). The geometric SD represents the typical ratio between an individual value and the overall geometric mean. Comparisons are adjusted for energy intake, and then for all potentially confounding factors using multiple linear regression. *P < 0.05, **P < 0.01. NE, niacin equivalents; RE, retinol equivalents.

Energy (MJ)9.32 (2.4)8.85 (2.8)1.05 (0.98, 1.13)1.05 (0.98, 1.13)
kcal2114 (656)2225 (563)    –    –
Protein (g)74 (20)69 (22)    –    –
% Energy13.5 (2.5)13.3 (2.9)1.03 (0.98, 1.07)1.03 (0.98, 1.07)
Carbohydrate (g)307 (89)295 (104)    –    –
% Energy52 (6.6)52 (7.0)0.99 (0.96, 1.02)0.99 (0.96, 1.02)
Fat (g)86 (27)81 (30)    –    –
% Energy35 (6.1)34 (6.4)1.01 (0.96, 1.06)1.01 (0.96, 1.06)
Thiamin (mg)1.6 (1.4)1.4 (1.6)1.09 (0.99, 1.20)1.08 (0.98, 1.19)
Riboflavin (mg)1.5 (1.5)1.3 (1.7)1.12 (1.02, 1.24)*1.17 (1.05, 1.30)**
Niacin (NE)33 (1.3)31 (1.5)1.03 (0.97, 1.09)1.03 (0.97, 1.10)
Vitamin B6 (mg)2.1 (1.4)2.0 (1.6)1.02 (0.94, 1.11)1.02 (0.94, 1.11)
Folate (μg)265 (1.5)237 (1.6)1.08 (0.97, 1.20)1.07 (0.96, 1.20)
Vitamin B12 (μg)4.6 (1.7)4.2 (2.0)1.04 (0.92, 1.18)1.07 (0.93, 1.22)
Vitamin A (RE)662 (1.9)572 (2.1)1.18 (1.00, 1.39)*1.19 (1.00, 1.41)
Vitamin C (mg)127 (2.1)115 (2.4)1.03 (0.86, 1.24)0.98 (0.82, 1.17)
Vitamin D (μg)1.5 (2.5)1.6 (3.3)0.88 (0.67, 1.15)0.87 (0.63, 1.20)
Vitamin E (mg)7.8 (1.6)7.4 (1.8)1.00 (0.89, 1.12)1.00 (0.89, 1.13)
Vitamin K (μg)17.0 (4.0)15.4 (3.1)1.03 (0.75, 1.42)1.06 (0.74, 1.52)
Potassium (mg)2796 (1.4)2621 (1.4)1.02 (0.96, 1.08)1.03 (0.97, 1.10)
Calcium (mg)850 (1.5)732 (1.6)1.11 (1.02, 1.21)*1.12 (1.02, 1.22)*
Magnesium (mg)238 (1.3)216 (1.4)1.05 (1.00, 1.11)1.06 (1.01, 1.12)*
Phosphorus (mg)1184 (1.3)1073 (1.4)1.06 (1.00, 1.11)*1.07 (1.02, 1.12)*
Iron (mg)10.4 (1.3)9.4 (1.5)1.06 (1.00, 1.12)1.07 (1.01, 1.13)*
Zinc (mg)8.1 (1.4)7.3 (1.5)1.06 (0.99, 1.13)1.08 (1.01, 1.16)*

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

This prospective study confirms the susceptibility of teenagers from low-income urban populations in the UK to the delivery of SGA infants, and demonstrates that this risk is not confined only to younger adolescents, but also to 16- to 18-year olds, who constitute the majority of pregnant teenagers. Previous studies have suggested continued maternal growth to be a primary cause of reduced birthweight.15 Yet, despite detecting skeletal growth in a third of subjects, we found no difference in the incidence of SGA birth between growing and non-growing teenagers. Instead, maternal growth was positively associated with birthweight, adjusting for constitutional and confounding factors, and higher rates of LGA birth, such that the distribution of birthweights in the growing teenage population resembled that expected for an adult population.

Studies of pregnant adolescents can be problematic because of their itinerant lifestyles and poor attendance at antenatal clinics, introducing difficulties with retention to the study. Despite these difficulties, this is the largest study to have investigated the impact of maternal growth in pregnancy, incorporating anthropometric, nutritional and endocrine assessments. Moreover, there were no identifiable differences in obstetric outcome or the demographic profile between those teenagers that attended for a second visit for the assessment of growth/anthropometrics, nor for those who donated a plasma sample for nutritional/endocrine analyses, and the whole population recruited to the study. Inclusion of an adult control group for knee-height measurements, and assessment of IGF-I concentrations as a biomarker for skeletal growth, strengthens our identification of teenagers as either growing or non-growing. In addition, our analysis is the first to use customised birthweight centiles, which account for maternal constitutional factors governing infant birthweight, and therefore provide a more accurate and clinically useful assessment of fetal growth than birthweight alone, especially in a still-growing teenage population.

The discrepancies between our findings and those of Scholl et al.15 may result from their use of unadjusted birthweights, ethnicity differences, or from the higher proportion of very young or multiparous teenagers in their study population. Multiparous teenagers may not have sufficiently replenished their nutritional reserves since their previous pregnancy, particularly with a short inter-pregnancy interval.34 This is consistent with higher rates of adverse outcomes observed in adolescent multiparae in a large UK population.35 Furthermore, studies that have suggested altered nutrient partitioning between mother and fetus drew from data collected 15–50 years ago, and therefore our findings may reflect higher BMIs in contemporary adolescent populations. Moreover, the teenage subjects in the current study had average gestational weight gains that exceeded those in the Camden studies.15 This hypothesis is supported by the observation of preferential maternal nutrient partitioning in chronically undernourished pregnant adults in developing countries, suggesting metabolic adaptations to favour fetal nutrient supply are dependent on adequate nutritional status prior to pregnancy.36 More recently, studies have demonstrated that teenagers in Mexico and Bangladesh cease growing to conserve nutrient supply for fetal growth.37,38 Our nutritional analyses support that growing teenagers have better nutritional status than non-growing teenagers. Together, these data highlight the importance of adequate maternal nutritional status in teenage pregnancies for both maternal and fetal growth.

In the current study, maternal growth was associated with greater mid-gestational weight gain as a result of a greater accrual of peripheral and central adipose stores. This was confirmed by elevated circulating concentrations of leptin, consistent with reports of an exaggerated leptin surge in growing teenagers.39 These data indicate a normal anabolic preparatory response in growing teenagers that ensures adequate maternal nutritional status in later pregnancy and lactation. We demonstrated a strong association between low gestational weight gain and SGA birth, and therefore the increased weight gain associated with maternal growth may be protective. Consistent with other studies of teenage pregnancy, the average gestational weight gains exceeded those recommended by the Institute of Medicine,3,37,40,41 particularly in the teenagers that continued to grow. The long-term consequences of excessive gestational weight gain in teenage pregnancy are a concern, and may contribute to the growing obesity problem in women of reproductive age in developed countries. Overall our findings are consistent with those reported by Stevens-Simon et al.,13,42 and suggest that pregnancy outcome in growing teenagers resembles that of adults, with infant birthweight reflecting maternal gestational weight gain.

Although we did not examine fat mobilization in late pregnancy, growing teenagers tended to deliver infants of normal-to-higher birthweight, indicating an adequate transfer of nutrients to the fetus. Moreover, growing teenagers had higher concentrations of IGF-I that, together with placental-derived hormones, induces maternal lipolysis to fuel maternal metabolism, elevates nutrient availability for maternal–fetal transfer, and promotes glucose and amino acid transfer across the placenta.21,43,44 We found no evidence of altered nutrient partitioning in growing pregnant teenagers; instead, the data suggests the hormonal milieu in growing teenagers promotes both maternal and fetal growth.

Our data implies that non-growing teenagers may be the more vulnerable group. They had a lower BMI during early pregnancy and poorer gestational weight gain. There is also evidence that maternal folate status may have been better in growers, although this difference was only significant in one of the three folate biomarkers used, and there were no differences in folate intake. However, as folate is an essential component of growth and development during all life stages, it is biologically plausible that insufficiency during pregnancy may affect the rate of maternal and fetal growth.45 We also provide evidence that the dietary intake of several micronutrients, including riboflavin, iron, zinc, calcium and phosphorus, was higher in growing teenagers. Both calcium and phosphorus are essential components of bone mineralisation, although the relatively high intakes observed in both cohorts were unlikely to lead to sub-optimal status, and therefore these results should be interpreted with caution. Some of these individuals may have ceased growing because of inadequate nutritional status, as observed in recent studies,37,38 whereas others may have already attained maximal height. We therefore speculate that skeletal growth may be enhanced in pregnancy in adequately nourished teenagers who retain the potential for growth. In childhood and adolescence, IGF-I, driven by growth hormone, is the primary mediator of skeletal growth. In pregnancy, the placenta secretes high concentrations of placental growth hormone, which closely resembles growth hormone in structure and function, and is a likely stimulator of the elevated IGF-I concentrations in pregnancy.46,47 As IGF-I was positively associated with gain in knee height, the elevated concentrations of IGF-I in pregnancy may promote growth, explaining the prevalence of growth in pregnant teenagers. Further studies are required to test the direction of causation in this relationship. These would need to assess the contribution of smoking, as although this is strongly associated with a higher risk of SGA birth, our data also suggest that it may impair maternal growth.

Maternal underweight in early pregnancy was strongly associated with SGA birth, supporting the hypothesis that poor maternal nutritional status compromises fetal growth. This is consistent with our concurrent detailed analysis of maternal nutritional status and pregnancy outcome in the same population, identifying in particular a strong association between poor maternal folate status and SGA birth in this cohort.30 Interestingly, obese teenagers were also at increased risk of an SGA birth. The etiology is unknown; however, studies of overfed adolescent sheep demonstrated that excessive maternal fat gain in early pregnancy was associated with impaired placental development and reduced birthweight.17 Whether placental development is altered in growing teenagers is unknown.

Our ability to detect differences in IGF-I and IGFBP-1 levels in non-fasting plasma samples between subgroups of teenagers (growing versus non-growing; SGA versus non-SGA births) is a novel observation. IGFBP-1 is classically negatively regulated by insulin; thus, our findings suggest that regulatory factors other than insulin are more influential in determining IGFBP-1 levels in the gestation studied. These may include known regulators, including IGF-I, or other pregnancy-associated hormones, such as progesterone or glucocorticoids.48 Interestingly, elevated IGFBP-3, in particular at 28–32 weeks of gestation, was associated with subsequent preterm birth. Whether this is because of reduced IGF-I bioavailability or IGF-I-independent effects is currently unknown. These findings add support to the potential for parameters of the IGF axis to act as screening tools for adverse pregnancy outcome. Further studies are necessary to validate their use in a clinical setting, and to explore potential differences in placental-derived stimulators of IGF-I and other metabolic hormones in teenage pregnancies.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

We conclude that maternal growth in otherwise healthy teenagers living in the UK is not detrimental to fetal growth. The greater weight gain and fat accrual in growing teenagers appears to support fetal growth, rather than hinder it. Moreover, growing teenagers have an endocrine status conducive to both maternal and fetal growth, providing biochemical evidence to support this hypothesis. Our data highlights the importance of adequate maternal nutritional status for optimal maternal and fetal growth during pregnancy. This study challenges some of the preconceptions about the causes of poor pregnancy outcome in teenagers, and may help to target improved antenatal care for those teenagers most at risk. Our findings cannot necessarily be extrapolated to teenage pregnancies in developing countries, where poverty, lack of education and limited access to contraception are likely to result in significant differences in pregnant teenage populations, and present distinct health challenges.

Contribution to authorship

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

The authors’ responsibilities were as follows: RLJ contributed to study design, analysed the anthropometric and endocrine data, and drafted the manuscript; HMSC contributed to the anthropometric and endocrine analyses; SJW contributed to the study design, analysed the nutritional data and assisted with drafting the manuscript; LP was co-principal investigator and contributed to the study design; CJH was the lead research midwife; PTS wrote the data analysis plan and conducted the statistical analyses; RO performed and contributed to the interpretation of the endocrine analyses; PNB (principal investigator) designed and co-ordinated the study. All authors contributed to the manuscript review and approved the final version.

Details of ethics approval

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

This study was approved by Central Manchester Regional Ethics Committee (#03/CM/32). Date of approval: 27 January 2003.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

This study was financially supported by the National Lottery Community Fund UK, Tommy’s the Baby Charity and Action Medical Research. We acknowledge support from the Manchester NIHR Biomedical Research Centre.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

We would like to acknowledge the assistance of Dr Melissa Westwood and Dr Martin Gibson in endocrine analyses and data interpretation, and Dr Sally Jewsbury, Dr Toli Onon, Ms Gemma Wild, Ms Gina Bennett, Ms Lorna Carruthers, Ms Dympna Tansinda and Ms Annette Briley for their invaluable roles in recruiting and performing measurements on teenage and adult subjects. Thanks are also due to the ATE Study Steering Committee (Suzi Leather, Chair; Lisa Bodnar; Jacqueline Wallace; Jane Sandall; and Mourad Seif). Above all, we gratefully acknowledge all the adolescents who participated in the study.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References